Shipping AI agents, without shipping risk

Most product teams building AI agents focus on getting them to work, but once they go live, new questions emerge fast. Which agent accessed what data, whose credentials did it use, and how do you prevent one user's agent from accessing another user's sensitive information?

This session walks through the security and governance challenges that surface when you ship customer-facing AI agents. Drawing from his experience as the CTO of Merge, Gil Feig will address credential management across multiple users, audit trails that actually work, and preventing data access misconfigurations before they become incidents.